Lattice basis reduction: Improved practical algorithms and solving subset sum problems (Q1340057)

A practical floating point \(L^3\)-algorithm, \(L^3FP\), is presented that shows good stability according to empirical tests up to dimension 125 with integer entries of bit length up to 300. Moreover, a practical algorithm for block Korkin-Zolotarev reduction is proposed, and a variant of the Lenstra-Lenstra-Lovasz \(L^3\)-algorithm is introduced that uses ``deep insertions''. For all the algorithms presented computational performance in solving subset sum problems is reported.